AI Model Enhances Cancer Prognosis and Risk Stratification

A new AI model, scSurvival, has been developed to predict cancer survival outcomes by analyzing single-cell tumor data. This tool identifies high-risk patients and links specific tumor cell populations to increased risk, improving upon traditional methods.

The model processes large-scale single-cell gene expression data, assigning weights to individual cells based on their relevance to survival. This approach preserves fine-grained cellular details osourcesen lost in broader analyses.

Testing on clinical data from over 150 melanoma and liver cancer patients demonstrated the model's accuracy in predicting outcomes. It also identified immune and tumor cell groups associated with varying survival rates and immunotherapy responses.

This development offers a more nuanced understanding of tumor behavior and treatment response at a cellular level. The research was supported by the National Institutes of Health (NIH) through multiple grants.